The present trial protocol is reported according to the CONsolidated Standards Of Reporting Trials (CONSORT) statement and Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) 36–38 (Figure 1 & Additional file 1).
The trial protocol describes a randomised controlled parallel-group superiority trial, the SmartBipolar trial, investigating the effect of smartphone-based monitoring and treatment in adult patients with bipolar disorder. The SmartBipolar trial is part of a larger CAG bipolar trial20, and therefore has some of the same secondary and tertiary outcome measures.
Trial design and study organization
The SmartBipolar trial is designed as a pragmatic randomised controlled parallel-group trial with a balanced allocation ratio (1:1:1) of adult patients with bipolar disorder stratified according to psychiatric centre from where the patients receive outpatient treatment as part of the CAG Bipolar specialised outpatient treatment (Psychiatric Center Copenhagen, Psychiatric Center North Zealand, Psychiatric Center Amager, Psychiatric Center Glostrup, Psychiatric Center Ballerup in the Capital Region of Denmark). The patients are randomised to the SmartBipolar trial during a six-month trial period (see description of intervention below).
The included patients are randomised to either active use of the entire Monsenso system, including clinical feedback (the intervention group) or to the use of the entire Monsenso system without clinical feedback (control group 1) or to mood monitoring only (control group 2). Patients continue with treatment as usual during the trial period. The flow diagram of the SmartBipolar trial is presented in Figure 2. The Smartbipolar trial is conducted at the psychiatric centers listed above. No changes in study design or methods have been made after trial commencement.
Participants and settings
All patients with a bipolar disorder diagnosis according to the International Classification of Diseases, version 10 (ICD-10) receiving treatment in the CAG Bipolar outpatient clinics in the Mental Health Services, Capital Region of Denmark, are invited to participate in the trial. The staff consists of specialists in psychiatry, psychologists, as well as nurses, social workers, and recovery mentors with a specific interest and clinical experience and knowledge about the diagnosis and treatment of bipolar disorder. The physicians follow evidence-based psychopharmacological treatment as according to established guidelines 39, and treatment also encompasses group-based psychoeducation, coordinated targets to improve quality of life by increasing concordance between clinicians, patients and relatives as well as continued ongoing supervision by a specialised clinical team 20.
Inclusion criteria: patients over the age of 18 years with a bipolar disorder diagnosis according to the International Classification of Diseases, version 10 (ICD-10) who are receiving treatment in the CAG Bipolar outpatient clinics in the Mental Health Services, Capital Region of Denmark are eligible for inclusion. Thus, the patients are not newly diagnosed, but receiving treatment for bipolar disorder during a more progressed state. Inclusion started in March 2021. Exclusion criteria: patients who don’t own and use a smartphone and who lack Danish language skills are excluded. Patients meeting the inclusion criteria and having none of the exclusion criteria are invited to be enrolled in the SmartBipolar trial.
Study procedure
All potential patients are invited to participate in the SmartBipolar trial by the staff from the CAG Bipolar outpatient clinics the Mental Health Services, Capital Region of Denmark. Since patients are recruited from the CAG bipolar clinics, all patients are diagnosed with bipolar disorder according to ICD-10 by highly trained clinicians with a specific expertise within bipolar disorder. Patients who accept to participate in the SmartBipolar trial are screened by trained clinicians to make sure they fulfil the criteria for participation. Following inclusion, the included patients are randomised separately according to psychiatric center (Psychiatric Center Copenhagen, Psychiatric Center North Zealand, Psychiatric Center Amager, Psychiatric Center Glostrup, Psychiatric Center Ballerup) to either the intervention group or one of the two control groups for a six-month trial period.
An overview of the data collection during the trial is presented in Figure 1. Information regarding psychiatric hospitalisation and duration of admissions, and the use of medication, will be obtained after completion of the RCT by linkage of the unique personal identification number (CPR), which is assigned to all 5.9 million persons living in Denmark with the Danish Psychiatric Central Register 40.
The intervention group
The authors have developed and tested a unique smartphone-based monitoring and treatment system, the Monsenso system for patients with bipolar disorder. The Monsenso solution was initially developed by the IT University in Copenhagen in close collaboration with the Copenhagen Affective Disorder Research Center (CADIC), Psychiatric Center Copenhagen, Mental Health Services, Capital Region of Denmark and a wide range of patients, clinicians and researchers (in the EU funded MONARCA project 28,29, the MONARCA II trial 30, and the RADMIS trial31). The effects of the Monsenso smartphone system in patients with bipolar disorder have so far been investigated in three RCTs by the authors 29–31. Overall, using the smartphone-based monitoring system resulted in increased quality of life, decreased perceived stress and increased use of Lithium without influencing the degree of depressive or manic symptoms.
Following the MONARCA studies and the RADMIS trial, adjustments to the Monsenso system were made, which now provides the following main features:
- Offer patients and clinicians across the Mental Health Services in the Capital Region of Denmark a tool for digital monitoring and support of treatment as a common way to substitute the general use of paper-based week schemes and clinical questionnaires.
- Make psychoeducation, personal treatment plans and cognitive behavioral therapy tools easily assessable via smartphone as self-help between clinical appointments.
- Collecting time series of data consisting of daily self-monitoring (mood, activity, sleep, irritability, mixed mood, cognitive problems, anxiety, alcohol consumption, stress, medicine consumption, menstruation, notes), clinical questionnaires as well as sensor data from smartphone aiming to:
- Provide an overview of clinical status, behavior, and adherence to and effect of medication and psychological therapy
- Use data to intensity research in CAG bipolar clinics
- Gain insight into psychopathology and treatment on sub-group levels (e.g., according to age and gender, bipolar subtype I or II)
- Optimize ongoing work with algorithms to predict relapse and episodes
The patients included in the SmartBipolar trial are invited to conduct smartphone-based evaluations in the Monsenso system daily. Patients in the intervention group and the first control group are invited to conduct evaluations of symptoms daily. Patients in the second control group are invited to conduct daily monitoring of mood only. In the two control groups, a study nurse without interfering in any way with the patients’ treatment will continuously monitor the adherence to smartphone-based self-monitoring and case of lacking self-monitoring, the patients will during the entire trial period be contacted and encouraged to fill in the data. Both the intervention group and the control groups are using their smartphones for the six-month trial period. In the SmartBipolar trial, the Monsenso system is available for smartphones capable of collecting sensor-based data on measures of illness activity on various levels (e.g., different levels according to type of smartphone - versions of Android smartphones and iPhones).
Subjective (self-monitored) measures of illness activity in the intervention group: The patients randomised to the intervention group, are prompted by an alarm in the Monsenso system at a self-chosen time during the day (standard 7pm but can be changed and also be de-activated by the patient) to evaluate subjective measures of illness activity dailys. The following subjective (self-monitored) measures of illness activity are available for evaluation: mood (scored from depressive to manic on a scale from −3, −2, −1, −0.5, 0, +0.5, +1, +2, +3), sleep duration (number of hours slept per night, measured in half-hour intervals), medicine intake (taken as prescribed/taken with changes/not taken), activity level (scored from very low to very high on a scale from −3, −2, −1, 0, +1, +2, +3), mixed mood (yes/no), irritability (scored from not present, present to some degree, present, present most of the time on a scale from 0, 1, 2,3 ,4), anxiety (scored from not present, present to some degree, present, present most of the time on a scale from 0, 1, 2,3 ,4), cognitive problems (scored from not present, present to some degree, present, present most of the time on a scale from 0, 1, 2,3 ,4), alcohol consumption (number of units consumed per day, 0 to +10 scale), stress (scored from not present, present to some degree, present, present most of the time on a scale from 0, 1, 2,3 ,4), menstruation for women (yes/no), a number (unlimited) of personal parameters (created by the patients themselves), and a free-text note. A patient can edit their responses in a self-assessment until midnight and the latest given response will be saved and shown in the web portal. If the patients wish to change their self-reports, they can enter a second evaluation in addition to the initial one, and both of the subjective evaluations are then visible for the patient and the health care provider when logging on to in the Monsenso system. If the patients forget to evaluate the subjective measures, they can evaluate retrospectively for up to two days. Pictures of the self-reported mood monitoring are presented in Figure 3 and 4.
Smartphone-based cognitive behavioural therapy (CBT) modules: Within the Monsenso system the following CBT modules are available: psychoeducation, behavioral activation, cognitive restructuring and rumination-focused CBT 41. The psychoeducation and mood-monitoring system include strategies for detecting and intervening with early signs of relapse. Patients are encouraged to use the module in an individualized and sequential way according to their condition, affective state, and insight and following guidance from the study nurse if wanted. The behavioral activation module includes activity monitoring and activity scheduling and focus on regulation of sleep and daily routines. The cognitive restructuring includes simple techniques to help patients to identify and modify dysfunctional automatic thoughts. A main purpose of the smartphone-based CBT modules was to help patients to reduce depressive ruminations. Therefore, a specific rumination-focused CBT strategy was included.
The clinical feedback: By offering all patients with bipolar disorder and clinicians to use the system, a unique world class data set is made, which can be used to optimize treatment to the benefit of patients and clinicians. Nevertheless, it is unclear and has never been investigated how smartphone-based treatment is best optimized - with or without a clinical feedback loop. The clinical feed-back loop includes 1) that clinicians continuously monitor self-reported patient smartphone data making early intervention possible when data suggest deterioration; 2) actively support and help patients employing the Monsenso tools including psychoeducation, personal treatment plans and cognitive behavioral therapy tools; 3) integration of the Monsenso system in group-based psychoeducation courses in the CAG bipolar clinics, and 4) meetings with the staff from the entire CAG bipolar clinic on a weekly basis where Monsenso registrations are monitored and discussed in collaboration (board meetings).
It is possible that patients will be able to use the Monsenso system entirely as a self-help tool in addition to treatment in the CAG bipolar clinics (see first control group below).
The control group
The first control group: Patients allocated to the first control group are invited to conduct daily smartphone-based monitoring without a clinical feedback loop. Furthermore, they will have full access to the smartphone-based CBT modules as described above.
The second control group: Patients allocated to the second control group are invited to conduct daily smartphone-based mood monitoring only.
Sensor-based data: Smartphones capable of collecting sensor-based data on measures of illness activity regardless of randomization group (intervention group or control group) collects as many sensor data daily during the trial period as possible based on patient given consent (iOS and Android: Battery level, Step count, Activity, Geolocation. Android: Light and screen state (on, off)). Thus, we will be able to investigate the use of sensor-based data for prediction analyses etc.
All included patients continue with their treatment at the CAG bipolar clinic. After the end of the SmartBipolar trial, patients will be offered the full version of the Monsenso system including clinical feedback.
Assessments
Since the SmartBipolar trial is a pragmatic trial and patients are recruited from the CAG bipolar clinics, all patients are diagnosed with bipolar disorder by highly trained clinicians with specific expertise within bipolar disorder. All outcomes, apart from registry-based measures, which are based on public registry-based data blinded for intervention, are based on patient evaluations (questionnaires and smartphone-based evaluations) and therefore assessed without blinding to the intervention. The patients are, regardless of randomization group, enrolled for a six-month trial period and invited to fill out questionnaires two times during follow-up – at baseline and after six months (Figure 1). Data collection of basic sociodemographic, clinical data and outcome measures (besides the primary and registry-based data) including questionnaires, will be handled electronically via the Redcap database in accordance with the standard procedure in the Capital Region of Denmark.
Outcomes
Primary outcomes
- Mood instability (based on daily smartphone-based measures of mood)
Secondary outcomes
- Patient-based questionnaires collected at baseline and after six months using the following questionnaires: quality of life according to WHO Quality of Life-BREF (WHOQoL) 42; self-rated depressive according to the Major Depressive Inventory (MDI) 43; self-rated manic symptoms according to the Altman Self-rating Scale for Mania 44; perceived stress according to Cohen’s Perceived stress scale (PSS) 45; satisfaction with care according to scores on the Verona Satisfaction Scale-Affective Disorder (VSS-A) 46. The validity of each questionnaire is available from the references. Patients will be contacted by the study nurse in case on non-response of the questionnaires. Furthermore, we will collect information on adherence to the Danish national guidelines of medical treatment of bipolar disorder according to use of the three main maintenance mood stabilizers for bipolar disorder: lithium, lamotrigine or quetiapine 39. Risk of psychiatric hospitalisation and cumulated number and duration of psychiatric hospitalisation according to data from the population-based Danish Psychiatric Central Research Register 40.
Tertiary outcomes (based on daily smartphone-based measures)
The following outcomes will be compared between the intervention group and the first control group.
- Mean stress (per month during the trial)
- Mean anxiety (per month during the trial)
- Mean irritability (per month during the trial)
- Mean activity (per month during the trial)
- Mean sleep duration (per month during the trial)
- Percentage of days with mixed mood
- Percentage of days with adherence to medication during the trial period
- Adherence to smartphone-based self-monitoring during the trial period
No changes in outcome measures have been made after trial commencement. All patients will be contacted by the study nurse during the study to ensure retention and complete follow-up.
Statistical power and sample size calculation
The statistical power and sample size were calculated using http://powerandsamplesize.com.
In the SmartBipolar trial with a follow-up period of six-month, 1/3 will be randomised to daily smartphone-based monitoring and treatment, including a clinical feedback loop (intervention group). 1/3 will be randomised to daily smartphone-based monitoring, without a clinical feedback loop (control group), and 1/3 will be randomised to daily smartphone-based mood monitoring only (control group).
Mood instability: For each patient, a mood instability measure based on smartphone-based daily mood evaluations will be estimated for each day and aggregated in accordance with prior definitions by applying the root mean square successive difference (rMSSD) method 13,14,47. The daily mood instability measures, reflecting the extent to which a daily and the previous day’s scores of self-monitored mood differed from one another during follow-up, were computed as the squared successive difference (SSD) of the reported values. Consequently, a daily mood instability measure could only be computed when consecutive daily values are present. The resulting SSD values will be aggregated for each patient as the root mean square successive difference (rMSSD), taking the square root of the mean of the SSD values 9,11,47,48. The instability measures will be computed following the original definition of the RMSSD score. Thus, differences between successive differences in mood are calculated, and then the values are squared, and the results are averaged before the square root of the total is obtained. The differences will be squared such that positive and negative differences do not cancel out when the values are aggregated by computing the mean. Squaring the values also puts more weight on larger differences. Finally, the square root of the mean puts the aggregated value back on the original scale. rMSSD is one of a few time-domain tools used to assess variability, the successive differences being neighboring mood scores. We decided to apply the RMSSD method following previous work. Thus, the mood instability measure, do not reflect the severity of self-monitored data, but rather the day-to-day changes by a statistical measure of a response to a question.
The primary outcome is mood instability between the intervention group and the two control groups – calculated as described above. Analyses will be done using intention-to -treat-analyses. According to recent analyses 15 mood instability in patients with bipolar disorder calculated using the definition above, has a mean of 0.80 (SD 1.93). If mood instability decreases 0.2 units (SD 0.53) on the mood instability scale for trial arm I (the intervention group) versus trial arm II (the first control group) and trial arm I versus trial arm III (the second control group), respectively, a drop-out rate of 10% during the six-month trial period (in accordance with the drop-out in our prior 6 months smartphone RCT[30, and a two-sided risk of type 1 error, α of 0.025 (since we will conduct two analyses between three groups with 0.05 level of significance in each analyses), a type 2 error risk, β, a statistical power of 80%, the sample size (N) is calculated as N =183. Thus, we need to allocate 67 patients in each of the tree arms of the trial and to randomize a total of 201 patients in the SmartBipolar trial.
Randomization
Sequence generation
The company randomize.net is used for the randomization (http://randomize.net). The randomization is conducted by the staff at each center using an online procedure with stratification according to the psychiatric center. Each center can log on to a webpage by using a secure code and will conduct the randomization on-site. Patients included in the trial are randomised with a balanced allocation ratio of 1:1:1 to 1) daily smartphone-based monitoring and treatment including a clinical feedback loop (intervention group) or 2) daily smartphone-based monitoring without a clinical feedback loop (control group) or 3) daily smartphone-based mood monitoring only (control group) (Figure 1). Block sizes are (unavailable to those who enroll participants or assign interventions) used to help preserve unpredictability 49,50. The study uses a stratified design, where patients are stratified according to the CAG bipolar clinic.
The statistical analyses will be adjusted for the stratification variable, and age and sex as possible prognostic variables. Further, in analyses on continuous variables, potential differences in baseline scores on the outcome in question will be included as a potential confounder. If there is no statistically significant main effects of age and sex, these variables will be excluded from the final statistical analyses.
Blinding
Owing to the type of intervention in the SmartBipolar RCT, the patient, the patient’s health care provider and the researchers are aware of the allocated randomization group. However, data on psychiatric hospitalisation can be assessed, blinded to the intervention status, by the researchers. The researchers are responsible for data entry, data analyses, interpretation of analyses, and writing of papers are kept blinded to allocation during the handling of data.
Statistical methods
Data from all randomised patients is collected until dropout or the end of the trial period. The analysis will be carried out with an intention-to-treat (ITT) approach. Concerning the primary, secondary and tertiary outcomes, analyses will be done employing linear mixed-effects regression models with a random intercept for each participant. Differences in outcomes between the intervention group and the control groups will be analyzed, firstly in an unadjusted model (except for potential differences in baseline values of the outcome variable in analyses) and then in models adjusted for the stratification variable CAG bipolar clinic, and for age and sex as possible prognostic variables. If there is no statistically significant main effects of age and sex, these variables will be excluded from the final analyses. Multiple regression models will be conducted investigating differences in the proportion of psychiatric hospitalisations and the cumulated duration of all psychiatric hospitalisations during the sex-month trial period. Models will be conducted as unadjusted and adjusted for CAG bipolar clinic (stratification variable), age, and sex. Missing data will be handled as missing at random.
For all statistical analyses, the statistical threshold for significance is p≤0.05 (two-tailed). Data will be managed by LVK and MFJ and entered using RedCap®. All analyses will be done using SPSS, version 22.0 (IBM, New York, NY, USA) and STATA version 13 (StataCorp LP, College Station, TX, USA).
Data about potential and enrolled participants will be stored in RedCap® and on secure servers at Monsenso during, and after the trial. Data collected during the course of the trial will be kept strictly confidential and only accessed by members of the trial team. Participants be allocated an individual trial identification number. Participant's details will be stored in a secure database. Only researcher from the trial team will access the data set. Potentially we will share anonymised trial data with other researchers to enable international prospective meta-analyses. This will be applied for separately.
Ethics, data permission, registration and dissemination
Ethical permission for the SmartBipolar trial was obtained from the Regional Ethics Committee in The Capital Region of Denmark and the data agency, Capital Region of Copenhagen (H-19067248, (P-2019-809)). The law on handling of personal data will be respected. The patients’ health care journals will be read to confirm information regarding the patients’ clinical history. The trial was registered at ClinicalTrials.gov as NCT0423042 January 2020. All positive, neutral and negative findings of the trial will be published according to the CONSORT guidelines 36. All potential participants are invited to receive information about the SmartBipolar trial on an individual basis by their mental health care professional where the information is given in a quiet and undisturbed room. All information is presented in both written and verbal form and participants can bring a friend or relative to the introduction conversation. Participants are informed that participation is voluntary, and that consent can be withdrawn at any time during the trial without this having any consequences for future treatment possibilities. All included patients sign a consent form and get a copy of this and their rights as a participant in a clinical trial. Patients do not receive economic compensation for participating in the SmartBipolar trial. This trial does not involve collecting biological specimens for storage. Results will be published in peer-reviewed academic journals, presented at scientific meetings, and disseminated to patient organisations and media outlets. The coordinating centre and trial steering committee consist of prof. Lars Vedel Kessing, Ass. Prof. Maria Faurholt-Jepsen, Malene Schwarz Dyreholt and Natacha Blauenfeldt Køster. Maria Faurholt-Jepsen, Malene Schwarz Dyreholt and Natacha Blauenfeldt Køster provides day to day support for the trial. Lars Vedel Kessing, Maria Faurholt-Jepsen, Malene Schwarz Dyreholt, and Natacha Blauenfeldt Køster are responsible for all aspects of local organisation including identifying potential recruits and taking consent. Lars Vedel Kessing and Maria Faurholt-Jepsen will be auditing trial conduct four times annually. Lars Vedel Kessing, Maria Faurholt-Jepsen and Nanna Iversen are supervising the trial and will meet four times annually. There are no Stakeholder and Public Involvement Groups. Interim analyses will not be performed. All potential positive, negative and neutral effect will be reported. Only the researchers will have access to the final dataset. All data including the smartphone-based data are owned by the researchers. Any data required to support the protocol can be supplied on request. The authors wish not to share smartphone-based data since they can easily be sensitive. Statistical codes can be shared upon request.
There is no data monitoring committee since this is a low-risk intervention being trialed. There is no evidence that suggests that serious adverse events would be anticipated. Potential minor adverse event may be anticipated. For this reason, we monitor worsening in severity of depressive and manic symptoms.